 Again, let me welcome you. Let me welcome everybody to the Future Trends forum. I'm delighted to see you here today We have a fantastic guest and a crucial topic. I'm really looking forward to our conversation Now we've been looking at open in various ways since the beginning of the forum We've been looking at open access and scholarly publication looking at open teaching We've been looking at open source software and demoing some of that And we've also been looking at open educational resources Indeed it's possible that we've been chronicling and forecasting a kind of ongoing wave of open innovation Indeed perhaps an open innovation Revolution where we switched to the majority of education being done in open ways Now one of the great luminaries in this field has been David Wiley A man who is so active. I feel he must be actually a conspiracy of three or four people at once Among other things he started off Lumen learning, which is a major enterprise for open education He's been a thought leader in this field and major activist and someone that we always turn to we want to learn where open is going I'm always fascinated by his insights and I'm just delighted to welcome here to join us. So without any further ado Let me bring out of the stage David Wiley Hello, sir. Hello. Thank you so much. Thanks for that introduction. Oh, my pleasure. My pleasure Remind us. Where are you coming from geographically today? I am in I just dropped in the chat box I'm in sunny, West Virginia today, not Western Virginia, but West Virginia, which is a different state Which I get confused about indeed indeed. Well, it's it's lovely here and we're just a couple hours east of you So I suspect Whatever you're getting we'll be getting very soon Yeah, it's sunny and kind of 75 ish here today Let's just cling to that let's just cling to that Well, thank you for deciding to be inside instead of outside like a sane person and talking with us David, what are you working on for the next year? What are the big projects and the big topics that are top of mind for you? Oh for the next year We'll the the main project I'll be focused on as a continuation of some work we started about two years ago that's an ongoing project to eliminate Race and income as predictors of student success in general education courses in US higher ed so Right now on the first day of class Before you even show up to class if I can find out something about you you or your family's income level or something about your race Or both of those things I can unfortunately make Way more accurate predictions about how you might do in the class then we would all like to have be the case so we've been for the last oh year and a half with some funding from the Gates Foundation working on New courseware model who's specific Actually whose sole design criterion is to eliminate race and income as predictors of student success and so We just piloted the first run at that this past January we had about I think we know about 75 sections across the country and an introduction to statistics Where the the platform features as well as the content itself are designed with this kind of specific goal In mind, so we just released intro to statistics in January like I said in the fall We'll be releasing intro to psych intro to business and quantitative reasoning And then we have several other courses kind of rolling out but all with this design goal How can we make it so that you know knowing anything about your race or your income level? just Becomes completely unpredictive right we just level the playing field and I I think the language of leveling the playing field might be more It's language that you hear more frequently But it doesn't have quite the sharp edge that I need it to have to understand if I see it and what I'm trying to do or not so very much a fan of this very specific language of Eliminating race and income as predictors of success. Yeah. Yeah, it's interesting It's almost perverse that you want at the end of this have you remove a reliable predictor Yeah, oh, that's excellent. That's like what's this project called so people can find it The the product I would say it's more product than a project the product is called lumen one Oh any Okay, and you can find it on our website, which is lumen learning calm You can you can find the introduction statistics course there now And the others coming in the future Excellent. Excellent. Well, I look forward to using that class as soon as I can There's a good demonstration of open learning and but also this kind of design. That's not terrific Yeah, I'm actually teaching intro to business this fall using the This release so be be eating the dog food here exactly. It'll be dog food in 101 Well, that's I mean that sounds like an enormous enterprise and a very beneficial one that could potentially scale up enormously And is that your is that your main project for the next year? Yes Everything else I'm doing is falling under that falling under that umbrella. That's it. I see that makes sense You know that the work we're doing around Kind of developing new models of student co-design Falls under that umbrella The work that I'm looking at doing an AI falls under that umbrella New work we're doing platform-wise in terms of enabling a B testing and things like that falls under that umbrella So there are a bunch of smaller pieces, but they're all kind of in the service of this broader goal Very good. Very good. Excellent. Excellent. Well, I mean, I have a feeling that right now People are just gonna be you know searching off on this and googling this and we're gonna have to fight to bring them back I think for a conversation, but but this is great. What a wonderful project Friends if you're new to the forum what I'm going to do now is just ask our guests a couple of questions To get the ball rolling and then I'm gonna step back and make space for all of you to ask your questions So as we speak think about what you'd like to put to David Wiley if you'd like to ask him about Loom in one or if you'd like to ask what about open that you'd like to raise. I just want to reiterate a point I made in my very quick introduction It seems you know, if we however far back we want to date the beginning of open education Maybe the 2000 maybe the Berlin declaration It seems like it's just been growing incrementally year by year. We have more more projects We've more more content a gradual kind of consensus that quality overall is improving We've had some back steps, but overall it's just been growing and growing. I'm curious Well, we'll be able to figure out when open becomes the majority status of education resources And when might we anticipate that to occur if possible? Yeah, that's a great question. I think the I think you're right that things have been Certainly growing in terms of the the amount of content and the number of people who are participating So there's been a lot of growth that way But you know reflecting on earlier today. I was thinking about the conversation we're gonna have today and I thought You know, I I published the first open license for content in 1998 and then published why Believe is the first academic book in 2000 that's published under an open license And then you look forward to things like just thinking about bigger scale things lot world knowledge in 2007 Open stacks in 2012 Pressbooks yeah sometime around 2017 There's a lot more participation and there's a lot more content But actually, I think it's sort of disappointing that there's not been a lot of progress made in terms of format You know the the kind of book that you would grab today from open snacks is Static words and text just like the book I published in 2023 years ago under an open license wasn't there's been There's been a real there's several Opinions in the in the OER community about the degree to which OER belong inside something like adaptive platforms or personalized homework systems or When you look at the technology when you think about the courseware technology account of the late Teens into the 20s and now as you start looking ahead at what's happening with generative AI There's been a large segment of the OER community that said that's not for us What we're about is about the resources that are free. They're completely free and only free And if they have to be completely free, then you certainly can't you have a Like a chat GPT tutor that's helping you along because you know that costs money per token to generate sponsors of support like that or or personalized courseware or adaptive courseware that has hosting costs and Data security and maintenance and other kind of costs associated with it so You just think about this morning. I was thinking about how this Real kind of zealous devotion to the resources being free In some parts of the community has stalled out the progress that we've made in terms of The effectiveness of the resources. We haven't been able to bring other tools to bear at scale Hmm because a lot of those tools have costs associated with them And there's you know some feeling that cost isn't compatible with the idea of OER So anyway, I was just reflecting on that earlier today I'm thinking we've made tons of progress in terms of the number of open textbooks that are available the number of people who are participating in creating open textbooks But at the end of the day 99% of what's getting created and called OER is still the same kind of words and pictures in a PDF or an Page that we're making at the end of the 90s Yeah, so a lot of progress in some ways and very little progress than others I would say. Oh, that's fascinating We've in a couple of our discussions about Recent developments in AI the idea is surface that perhaps instead of or alongside of the black box AI produced by enormous capital intensive giants like Microsoft and and Google instead we should have open source alternatives or liberal alternatives That also then use for data sets open or open licensed content Do you do you think that we is that in our future? And of course we're talking about AI future We're often talking about like one week But I mean, do you do you think that's something we should anticipate over the next year or so? Over the next year or so. Yeah, I mean so much is gonna happen in The next week like you said today, right? Just today or yesterday Facebook released a paper detailing a new model that looks to be Significantly more parallelizable than transformers Quite a bit faster You know, there are people proposing new models all the time, but this one out of Facebook I think it's megabyte is actually the name of the mall And you should be able to find a reference to the paper on you know on Twitter or somewhere But there's just it feels like every day is a year right now You know with what's happening in this space the Once these models bring down the cost of compute then there will be a lot more participation from You know open community But when it takes you know when it takes months and fifty million dollars to Train a single base model. You just you're not gonna have a lot of you know A lot of kind of community coming together to to do that but I Think that all the indications are That you know hardware is going to continue to get better the models are going to continue to be more efficient Maybe be parallelized like this new megabyte model from Facebook It would be great if at some point well Let's actually talk about this for a second. I started to say it'd be great if in the if at some point in the future There'd be base models that were fully open But I think where the action really is going to be at least for us in education is in the work We're going to do kind of layering value on top of the base models through fine-tuning or whether it's through prompt engineering or kind of Injecting additional information to prompts using embeddings or something like that Expensive to develop that base model and to some degree It's not clear to me whether we Need a fully open base model or not because Use going to be the educational value is going to be added in the layers above that in the fine-tuning If it turns out that that's necessary or in the embeddings and the prompt engineering Because even if you have a model that's fully open If the data were all open and the code were all open The way it operates is just as black box-ish as it is if you if you can't see the code You know the way that inference happens the token prediction happens and It's still cost money It's got to be running somewhere You know every time someone asked a question that's going to cost money to generate an answer So even the vet was fully open from front to back you there'd still be costs associated with using it And I just it seems like those base models are already so commodified if you jump on hugging face there are Dozens maybe hundreds of base models that are already there, so Maybe at some point there'll be some as this starts to slow down if it ever does Maybe there can be some consensus around the base model. That's kind of most useful for our Use case here in in education now. I say use case like there's a singular one But you know everything I think our minds all immediately go to tutoring that's But just imagine imagine if you have an advisor an Academic advisor who actually knew all the policies actually knew Actually knew everything about every course you needed to take and what see things they could be taken in I just think there's so many So many applications across the space for us You just said a lot in a few words No, no, that's why why we brought you here and we had a couple of quick questions One is a people asked what an LLM is and that's a large language model No, it's okay. I just want to make sure everyone feels that feels comfortable We have a large language model, which is a recent version of AI which we could talk about if you like a bigger question was What is a base model? How do you distinguish that from what's on top of it? Oh? so the in when you think about a jeep the GPT the acronym you're talking about kind of a the If you think about the the difference between GPT for and chat GPT All right, so GPT for is a is a base model It's a model that's not trained to do really much of anything except predict Tokens that should come after each other and so there's an idea with In the in this space of these kinds of transformers that they get pre pre-trained in a very generic kind of sense and then there's going to be some fine-tuning that's done afterward Like chat jeep the difference between GPT for and chat GPT is that chat GPT has been fine-tuned to have a Conversational kind of interface where you can ask and it can remember and respond and you can carry on a conversation as opposed to More of a one-shot kind of interaction where you ask it one thing and you get a response back Well like with GPT for and then if you were to ask it to tell me more about that It would have no idea what you're talking about it comes it it hasn't been It hasn't been fine-tuned to perform that come in that conversational sort of way So the idea of the base model is just you know bringing in tons of data Analyzing it being able to do some predictions of you know, what character or what token ought to come next as you're stringing together Response to a question or an answer to a prompt like write me a poem or whatever that might be But the base model Is good is a good kind of general purpose Model, but we don't tend to interact directly with the base models We tend to interact with ones that are fine-tuned like chat GPT for a specific purpose or like con academies Con me go which has been a specific purpose of being you know Playing that tutorial kind of role To use an example that will resonate with no one and I don't know why If you ever participated when you're a young if you ever built the cut the Pinewood Derby, yeah, you know, you had this block of wood And then you would nail wheels on to it or not or paint it or shave it down or do whatever to it like think of the base now it's just being that kind of block of wood and You need to put you need to put the wheels on it You need to put the weights on it. You need to paint it. You need to cut it down for the stickers on to your yeah stickers For your specific purpose well so far out of 65 people here two people have found that no three have found that to be a great Knowledge, so it's not it's not nobody. This this is good Fred I'm dating myself here. Oh, no, and I'm clearly just sprouting more white hairs right here The well first of all, thank you for that excellent excellent answer to to that question and speaking of questions we have a bunch of them coming up and some of them refer to the Project a lumen one and I want to make sure that everyone got a chance to to put their questions out about About this and friends. We have more questions coming up as well So this is one from our dear friend Phil long who says what are the variables influence leveling the playing field? Is there any aspect of selective disclosure or is this environment asynchronous? Oh Well, I would say that the hardest part about leveling the playing field is Let me talk about the opposite the thing. That's the easiest to do and that In my I'm sure this is true for many of us on call in in my career done a couple of times is See you've got some gap in performance between the group that you're trying to bring up to To the par with the other groups in your class It's really easy to intervene in ways that improve everyone's performance But all that does is slide the gap to the right You see I mean so Improve the performance of underperforming students but that same intervention also Improves the performance of who the people who are previously high performing students So now we've just we've slid the whole gap upward instead of closing the gap Right, so what kind of interventions really can have an outsized role on students who are underperforming? And help close that gap instead of shifting the just moving everybody You know up 3% or moving everybody up half a standard deviation or what? and That turns out to be I Mean just improving performance in general is really hard Improving it in this really targeted way turns out to be even harder, but it has to do The I would say the the success that we've been having With that Has to do with really kind of engaging students engaging those students who? Who need the most support and involving them directly in the design process? For the course materials so that we're not designing for a generic student. We're designing for Right, you can't you can't design a new book for every one student in the context We're working in now now the generative AI opens up some different possibilities But for example in the introduction to statistics course You can easily imagine if you can remember back to the time if you've ever had a statistics class There's just so many example data sets that you encounter during statistics where you would find the average height of Adult males and the average height of adult females and they compare them to see if they're statistically significantly different and it's kind of hard to imagine an example that would be less engaging or less interesting than the average height of people and yet It comes up kind of all the time So in talking with students You know About the design of this course and kind of the data that we're going into it and the kinds of things They're gonna be asked to do in the way they're asked to work You know, we heard a couple of messages loud and clear which was we're not gonna engage if it's boring We're not getting engaged if we don't care or we'll minimally engage Those examples that you're talking about are awful like great. Give us better example Give me a better example and we'll go get it. Okay. Well, how about Spotify streaming data? Like who are the top streamers for 2022? Those are data that I care a little more about and are a little more interested And don't ever I mean while we're at it don't ever just give me one data set Give me two or three or four or five data sets so I can choose from whenever I'm gonna do some analysis because I Want to work on something that's engaging to me and interesting to me You know another thing we heard from students was All the time it happens that I work through whatever I'm supposed to work through and I listen to the faculty member talk And I still don't understand so I just go hit YouTube and I watch random YouTube videos until I find somebody that You know that I can kind of relate to and who I understand But then every time I do that I get lost and I wander off down these trails as we all do you know when we Go out into the internet you Recommended video and then three videos later. You can't remember how you started so You know another thing that we've done in lumen one that has been That there's been strong feedback from students about is don't just give me recordings of a single faculty member For every topic that's going to be explained or every problem that's going to be worked in a video Have there be two or three or four faculty members and have One of them be black and have one of them be Hispanic and have one of them be white I'm one of them more like this and another one more like that and give me real choice So that I can find the person who I relate to that they're all working exactly the same problem teaching exactly the same topic You're gonna cover it, you know in a little different way and talk about it with this kind of example as opposed to that example So really having students involved Not like creating all the content creating the course design creating all the features For the platform and then asking the students what they think about it, but really having them involved throughout the whole process Has led us to do some different things that are Having some impact Well, that's great. I sorry just personally. That's terrific to have students play a constructive co-creative role Why is it broccoli in the first place and I realize I just need to ignore the chat box because I'm not keeping up I It is and and we will have a good question from that from that brilliant guy in just in just a bit But yes, there's nothing quite like the surrealism of appearing in mid-air is But we're following up on this we have a question from our dear friend Steve airman who wants to join us on stage So let's bring him up on the screen And Steve is coming to us. I think from Maryland. Hello, Steve. Hi, Brian. Can you hear me perfectly great to see you? David I had two comments about different things that you were talking about the first one was tutoring I Was Had the good luck to be able to sit down in a session of a course full undergraduate course where Undergraduates were learning how to be learning assistants In the in some course that they'd already taken usually a STEM course and what I wanted to point out particularly was they were learning a how to build a mental model of what the students they were tutoring How they were thinking about this and then using that mental model to try Both testing it and maybe advancing the students by asking the students something or Yeah, we'll ask in the student to do something typically what it was So in my mind tutoring especially the kind that's gotten such You know, I multiple sigmas or whatever it is I think the tutor is always making a mental model of where the student is And that's as I understand it fundamentally different from LLMs And I so I think I have a problem with applying the same word to two things that are so completely different from each other Oh, well, I would say that they're different from each other if you leave them alone but say that you were say that you're me and the primary context that you're working in is a Adaptive courseware platform where you have in your software you have a model of what the students understand them what they know Where they are along each of the specific, you know outcomes or concepts in terms of mastery And then when I go to launch a student when I say when they get to the end and they say I'm still confused Let me talk to the tutor when they launch that tutor The first thing I'm gonna do is I'm gonna send the tutor my whole model of where they are in their understanding So that when chat GPT or whatever software it is begins that conversation. It's gonna be before it even says hello It's gonna have from me my estimates of where the student is in terms of their mastery of the different concepts and it's going to be I don't know how much you've looked into prompt engineering, but you can Scaffold the kinds of interactions that the agent is going to have With the student you can tell it things like don't just give them the answer Prompt them for what's the next step that you you can do quite a bit of scaffolding And then you can augment that scaffolding with additional context around Where students are in their current understanding what they have mastered what they haven't mastered And I mean we're all single-digit number of months Into trying to do this work But there is there are ways to push context in back into the LLM as it's Like we'll just let's just say chat GPT But understand that I mean any large language model when when I say that there are ways that you can feed context about your Understanding of what's interesting to the students what the student is understanding what they're not understanding You can feed all that context to it in a way that it will be responsive to as it engages in a tutoring session with student Yeah, I'll just say briefly because I did want to bring up a different issue I think there's still more steps to take in developing that kind of model that are based on student misconceptions for example Anyway, the other thing I wanted to talk about was I'll be trying to be as brief as possible Closing the gap as you put it and I would say while raising the playing field This is a subject that a lot of people have been working on for 20 years and more and To encapsulate what I understand it high-impact practices are have been shown to be very important in raising the Playing field while also narrowing the gap Learning pathways It goes also not just to educational strategies and I could have listed about five more But also to organizational changes For example the infrastructure that would help manage Open learning in a variety of senses online the slash open and I think It's sensible to start by saying In principle, these are the kinds of things that need to be happening In order for the education to be better for everybody to close the gaps and to be affordable At the same time you need to be thinking about all three all three targets and then after you have You know how much of a model you want to build then ask how can OERs help this model work better? And I think it provides maybe a different Slant on some of these things or at least a way to say these the OER has become effective as Part of a larger constellation of changes. I could not Possibly agree with you more I think another thing that is to some degree holding back the OER community is At at some somewhere along the way OER stopped being a means to an end and became the end Where instead of saying I have this pedagogical approach that I'm using OER fits into it in this way like as I'm reaching into my belt of tools I'm gonna pull out the OER tool now because I'm doing something that calls for that to the point where just We just celebrate whenever anyone adopts OER without even looking to see Did it make an impact on student learning? Did it hurt student learning? Did it help student learning? I mean the the the empirical studies are It's just what you'd expect right? It's no significant difference is the vast majority of stuff in the middle and out on the end You have some that there's some that show improvements and outcomes and some that show You know that outcomes suffer but I'm kind of I'm a little confused at at this point now and I admit ten years ago I was caught up in it myself and it's taken me a little distance to get to where I am now But it's like celebrating the fact that somebody adopted one textbook over another it It's not all the things we need to be thinking about maybe it's great It's 100% amazing saving students money is So wonderful, and if you can say if you can get the same outcomes For a fewer dollars per outcome Yeah, then we'd call that a win normally right but then when you go back in and look at outcomes when you look at the four-year graduation rate in a hundred fifty percent time the way the Department of Ed tracks it How many what percentage of students complete their four-year degree in six years? What percentage of community college students complete their two-year degree within three years that number is about 30 percent? We made it more affordable for 70% of people to never complete their degree Ouch making it more affordable can't be the end goal and Yeah, I one of the things I think that's holding that community back Is that that saving money has become the goal? and so our eyes are kind of off the ball of The thing we ought to actually be caring about which is student learning student outcomes Completion, you know the effect of completion on their life after school things like that, so I I I think that's what you were saying Steve and if it wasn't I apologize, but I I can't sing that song Frequently enough or loudly enough so amen to everything with that whole line of thought Well, thank you. Thank you Steve for starting the choir going And everyone needs to read Steven's book as well as to look at his appearance in the forum because they're both excellent And David, thank you for the excellent excellent answer Friends if you're if you're new to the forum, you just got two examples one of a Q&A question and one of a video question So please feel free again that white strip in the bottom to press either button for whichever is his best for you David I want to clarify and redeem the name of the broccoli asker Which is a sentence nobody on earth has ever said before so I feel pretty proud of that And this is we have a typically great question from Tom Hames Who was mentioning that I just want to put that on the screen for you here How much does a very concept of going to class have to do with success based on economic class? Which is closely tied to race isn't the system stacked against you? The system is stacked against you mean some of these issues are systemic, right? And there's actually a paper that I just saw yesterday if you'll give me just a second I will pull up the reference and drop it short in the in the chat here Making the claim this is a study with 28,000 students in it and here's the link in the chat And here's the kind of here's the key takeaway Class attendance is a better predictor of college grades than any other known predictor of academic Performance including scores on standardized admissions tests like SAT High school GPA study habits or study skills Class attendance is a better predictor of success than any of those things now some of those things are just well terrible predictors So it's not a very high bar SAT is just not a great predictor But just you asking about class attendance. Just made me, you know, think of think of that question from yesterday, but Yes, there are there are absolutely structural problems that keep people From especially particularly physically from getting to class, right? Can you put the question back up Brian, I feel like I oh sure of course of course It's a deep one. How much does a very concept of going to class have to do with success based on economic Class, which is closely tied to race and isn't the system stacked against you. Yeah That's a yes or no question, so I'm just gonna say yes It it it's stacked against you in so many ways and We we have a chart that we use when we onboard new employees at Lumen and it's a pie chart and About ninety nine point five percent of the chart is one color and then a very narrow sliver of it is another color well and The the message that goes with that chart is here are all the things that need to happen to be able to Dramatically improve student success with very high probability And then there's an arrow that points to the little sliver that says this is the part that we work on. Oh There there's food insecurity. There's housing insecurity There I mean there's just so much that students are swimming upstream against part of it is You know part of it is the cost of their course materials and a bigger part of it is the effectiveness of their course materials and the degree to which those materials You know implement best practices and are really gonna effectively support them in their learning but You know, but we we use this kind of exaggerated pie chart just to make the point that As a like course material provider, there's so much we can do but there's only so much we can do right It's why it's amazing that there are things like the fast fund out in the world to give people There's a whole ecosystem of problems and there has to be a whole ecosystem of solution providers and You know this project that we're working on now We're trying to eliminate race and income of predictors of student success I'm very curious to see how much of that gap we can close I would love to imagine us completely closing it, but there are a lot of other things that Structural things that impede students so we'll see how close we can come to it You know a funny side note here Is that when we receive this grant from the Gates Foundation to build a brand new platform adopt and adapt OER into that platform Etc. It's all the things we were funded to do There's a separate grant given to digital promise to run a nationwide RCT to evaluate the effectiveness of the thing that we built For the thing that we're in the process of building and it just goes to show How incredibly difficult it is to run kind of nation-level? RCT type of research that the grant to digital promise is actually bigger than the grant to Lumen learning was Wow Building a brand new platform all this kind of right doing all these things You know paying these groups of students at institutions around the country to participate in these co-design sessions all the things we're doing It actually costs more just to run the RCT To figure out did we or did we not or to what degree did we actually manage to eliminate race and income as predictors of student success? Well, well this also speaks well to how to how effective and efficient the Lumen is I would say Thank you Yeah, there's there are There's a primary benefit of using OER and then there are multiple kind of secondary benefits to using OER But you know dramatic increase in improvements and efficiency I would say is definitely one of the one of the benefits of using OER to my mind The primary benefit of using OER is that it makes it possible for you to engage in continuous improvement Cycle after cycle based on data that students and faculty generate in their use of the courseware so say For to make it make it easy imagine that there are 15 modules in this Courseware and in one of these modules I can see that students are consistently engaged With all the interactive practice They're looking at the feedback. They're going back and practicing again. They're watching the videos. They're doing the homework I've interviewed their faculty member, and I know their faculty member is doing the in-class Active learning activities that we that are prepared and that come as supplemental materials, etc Students are doing all the things we're asking them to do and they're still not learning Or they're not learning at the level we want them to that tells me that our materials aren't good enough And I need to improve them And so if the materials I've adopted are OER I have permission to you can say revise or remix or however You want to think about it? I have permission to make the changes to those materials It can help them be more effective next semester than they were this semester and that to me Seems like it's the most important reason At least from limits perspective. That's the primary reason that we use OER We love the fact that it helps us keep the cost down as well Because there aren't royalties to pay and things like that that a traditional publisher would have to to deal with But I actually think to continue some improvement part of it's more powerful But you've got uh Well, thank you. Thank you. That's that that is so powerful and I can think of some stories where that's incurred And by the way, this is just a shout out to Alan Levine who is on this call and Alan has a Oh Yes, indeed. He's an ongoing project of Great stories about success using open. So just a shout out to him a great person Uh, we have more questions coming up But I want to make sure that that everyone gets a chance to handle them and these these now focus more on an OER per se and here's one that's coming up from Ray Garcelon So what do you think are the current barriers major barriers to OER adoption for faculty institutions? And are they different for the major barriers from five to ten years ago? Um, yes, they are different. I think five or ten years ago The majority of faculty never heard of OER and you can't adopt something if you've never heard of it um, I think now more people have heard of it and I would say that now the major barriers to adoption are Um Say I teach college algebra and I currently use Whatever product pick a product. It's in the 150 to 200 dollar range But the main thing that it does for me is it automatically grades student homework and it saves me From all weekend long hand grading all the homework problems that all my math students have done And then you walk through the door and say hey There's this thing called OER. It's amazing. You get five hours. There's open pedagogy. There's all this stuff um And it's going to save your students a ton of money But now instead of them paying you're going to pay because you've got to go back to hand grading everything you were doing before Well, that's just That's not an adoption pitch that is ever going to work So you have to put OER together with These homework systems or some kind of adaptive platform or something like that faculty don't want to move backward in You know the support that they're receiving currently from traditional teachers They're getting quiz banks. They're getting PowerPoint slides. They're getting pacing guides. They're getting they're getting all this material plus support for grading plus In some cases they they might even be getting like some dashboards that are showing them marginally useful data giving them some insights about what's happening in the class and if the pitch is Stop doing all of that stop having all that done for you and start doing it all by hand And you can adopt, you know, this OER then People just won't the the vast majority of people won't do it. Yeah. Yeah So, you know, the reason that lumen exists is because we came out of a two-year grant back in What this has been from 2010 to 2012 Where we uh, my friend kim who's my co-founder and I and some other people were trying to support faculty in making this transition From tc from traditionally copyrighted materials to OER TCM being the opposite of OER And what we heard over and over and over again was this is too Hard because publishers used to do all these things for me Now you're telling me that I have to do all these things for myself and so You know part of part of the reason that lumen exists is we realize that if faculty were ever going to adopt OER Somebody's going to have to do the work of providing all that supplementary supplementary material banks Powerpoints in class active learning activities blah blah blah plus the platform that can automatically grade the homework and can post the grades Learning management system and all those other things But it doesn't need to cost $175 Right it costs. It's going to cost some money in lumen's case. It costs $35 which I think is a steal, but i'm biased clearly um but And I think that you know coming back to the point that uh, we were talking about at the beginning if your view of OER is that OER in its OER needs to remain pure in its unadulterated form and be unaffiliated or unassociated with anything that has costs then You really are Swim upstream trying to tell somebody who's used You know a pierce in my lab for the last six years and now switch to this PDF that's free for your students But it comes with no support for you so I think the barrier to adoption now is Is that finally getting back to the the question right is providing those supports and tools and The dashboarding and the other things that faculty have come to rely on and Thankfully actually started to make some use of finally Oh, that's a that's a great answer Daunting answer ray. Thank you for the great question. It reminds me I asked the head of the company that owns Elsevier the scholarly publisher If if what business he thinks he's in and he says he's in the data business not in the pdf business Which which was very very telling We have more questions in only seven minutes to go So I want to make sure we all get a chance to pop these off From Giselle LaRosse who I think hopefully is coming to us from New Orleans today She asks for me. It seems that what's missing in the context for diy learning or homeschooling for adults What if we are as a front runner to larger systemic change? um I think that's so as a person who for some number of years in our family We've homeschooled all of our children for some number of years before they've transitioned Into either a brick and mortar or a or an online I'm super sympathetic to this kind of view and I do think it's possible that I wouldn't think I would I would say it a little differently I would say and I saw cable's name blow past in the chat there a minute ago cable had a dime for every time he heard me say the seat could retire I think that content Is infrastructure Right and so when When you're going to create A new course when you're going to create a new Learning tool when you're going to create a large language model when whatever you're going to make um It builds on top of content and you can go Acquire very expensive content that you have absolutely no rights to and try to build things very in a very brittle way on top of that Or you can acquire open content that lets you do all the things that you need to be able to do and really Helps you get faster further less expensively And really importantly when the infrastructure is open it lowers the cost of experimenting Right so more people can try more things more and And we can't depend on you know the four or five Whether it's the four or five biggest Software companies that are creating a i tools or it's the four or five biggest publishers that are creating content You just can't depend on four or five big anything To solve any problem. That's not how it's going to work the solutions are going to come When they're or they're going to come work dramatically more quickly when there are thousands of experiments running in parallel Right, that's how you're going to you're not going to find solution by doing five things You're going to find it by doing a thousand things. So how can we make it less expensive less risky? Faster to engage in those experiments. There's a great quote, which I will Mangle a little bit don't get it close enough that the that people can find it if they're interested honest Torvalds once said about the Linux operating system and So don't ever make the mistake of thinking that you can design something better Than what you get for massively parallel trial and error with a feedback loop That's getting your intelligence way too much credit We we need this massively parallel experimentation trial and error. We need yes people You know engaging all these experiments and when the content That you want to do that with is freely available to you to try the thing that you want to try Then it's a lot easier to go and do that experiment that you wanted to do Well put well put and I'm glad that you can do some of the work at scale through lumen But we only have a couple minutes left. We have another question that comes in And this is from Ed Finn while prezzo he asks about the digital divide Do you feel the digital divide adds another layer to reinforcing so as to recognize barriers for students While OERs may be free or low cost if they can't access them. Does that matter? Well you can you leave it on the screen, of course, yeah, give me give me one second I'll extend the time which not knowing shindig is is actually is the question I meant to ask. Can you leave it On the screen. Yes. I can yeah um I mean the digital divide of course adds a layer here that is That is reinforcing some of those barriers, but there are people You know lumen doesn't do this work like I talked about a minute ago There are lots of problems that students are facing. There are lots of people working on solutions to them I don't work on the digital divide problem, but there are people who do and I love the fact that they do and they They're doing good work I think we all wish that that work were coming along faster That everybody's everybody's doing the best they can Of course, if you can't access the materials they benefit you exactly zero So this problem of Providing access like do you really need a phone first experience because the majority of students that you care about are on the phone Maybe don't have a laptop or don't have Certainly don't have a desktop or and they're not able to stick around the lab long enough on campus or Um, you know, what's that? What does a phone first experience look like and that's not an experience that you know major publishers are Are building right now. So those are you know Something that is if not phone first is at least highly responsive is the kind of thing that you would do if you were trying to say Eliminate race and income as predictors of student success Well, that's a good point. Thank you for your clarity and and directness there I have one last question here from the from my side before the clock completely runs out Which is uh, what are some of the emerging Forms that open may take or may enable if we all start taking open more seriously Because if we go beyond what you said, okay, we save students money. Yes, this is good If we go beyond that, um, do you think we should see more and more projects like lumen one? Do you think we'll see more that kind of distributed feed parallel feedback loop? Uh, what are some of the other, um, new things we should be expecting? Well, really, I think the the feedback loop part of it, uh, that you mentioned is really the main kicker um, there educational research is so full of frameworks and theories and models and All of which are synonyms for opinions that have never been empirically validated It might sound hard to believe, but it really is true. I was taught so many instructional design models as a graduate student that Why would I use this one instead of that one? Well, because that one's maryls instead of wrigoluths and Well, but when we create instruction using those two different models and put them in the field Who learns more? Why wouldn't you do that kind of? You know, like there's almost like a and just a lack of curiosity around empirically validating the work that we're doing so um You know, I mentioned ab testing very quickly earlier Maybe make let me make another little plug for it here to the degree that a lot of the things that we're doing are digital It becomes um It becomes pretty easy to employ some of the tooling That honestly we would borrow primarily from like the e-commerce or the digital marketing space So think about all the tools think about all the things that amazon does in order to maximize The the dollar sign next to your shopping cart before you check out Now think if you were to apply all those tools in instructional context And what you're trying to do is maximize the number there Next to your name in the gradebook on that exam that you just took Okay, so it's an optimization problem We're testing lots of things we're and ab test is just another name for a randomized controlled trial It's just an rct That's the gold standard in education research according to the department of education is the rct So much of this happens digitally Why aren't we more curious about? How well things work At all and then when we find that something's maybe not working very well instead of just blindly swapping it out with something else Why not also ab test that new thing and see if it's actually an improvement over the one you had been using before I wonder if it's because so many of the educators and economics involved are from the humanities And this kind of social science approach is often not part of their lives Yeah, yeah, yeah, and I should say too. Hey bfa music. So I'm I'm I'm I'm admittedly a little weird in that regard, but um, but I do think that there's a There's just there's not a culture of There's not a culture of validating. There's not a culture of Even though even with the emergence of the idea of the kind of scholarship of teaching and learning You'll only see that really taking root in a couple of disciplines. You don't see that kind of happening broadly Yeah, so I think if people if we could help faculty be more curious About what's working for their students and what's not And feel some ownership over that themselves I just Would unlock so many things and oer would be so useful in the kinds of things that those faculty would do once they had that Mindset. Yeah, but giving them the oer and not helping them develop that mindset just gets you cheaper textbooks that You know are gonna result in the same kind of poor outcomes that we're getting right now So maybe that's one of the revolutions we can expect is that kind of you know deeper curiosity as you said and awareness of Of how these tools actually assist or don't assist learning Yeah, uh, David, this has been fantastic a blizzard of an hour with so many ideas What what's the best way to keep up with you these days? Is it through open content on twitter or to the luma website or where? um Yeah, either open content on twitter or my linkedin. I generally cross post um But yeah, either of those would be would be fun. Very good. Well, thank you for being with us Thank you for all the work you do and we look forward to hearing what uh, what the digital promise thinks about what lumen one has done Yeah, check check back in the spring of 2026 and You know, we'll have an answer But I want to thank you too. Thank you for the invitation. This has been super fun to see so many folks I haven't seen in such a long time and Under pleasure it's been fun. Thank you so much. Oh, thank you. Thank you. But don't go away yet friends I just want to first of all second that David's delight in all of you. I completely feel that myself If you want to keep talking about these issues again, you can at open content on twitter and at me brine Alexander Or you can hit me up on mastodon. There's my long URL and just use the hashtag f t t e If you'd like to look into our previous sessions where we've covered some of these topics including ai including open education Including the educational change. Just go to tinyurl.com slash f t f archive If you'd like to look at our upcoming sessions Which do touch on assessment as well as college teaching and teaching with ai and campus economics and war Just go to forum that future of education that us And above all, thank you all for thinking and sharing so much today. It's been a real pleasure I hope all of you are heading into june with a lot of happiness and delight I hope all of you are safe and sound and we'll see you next time online. Take care Bye. Bye